[torqueusers] Bizzare $PBS_O_WORKDIR bug

Garrick Staples garrick at usc.edu
Tue Sep 4 01:05:55 MDT 2007


On Tue, Sep 04, 2007 at 03:22:03PM +1000, Chris Samuel alleged:
> OK - this had me completely stumped for a while trying to work out why 
> certain of my test jobs would fail completely.
> 
> In Torque 2.1.8 it appears that the $PBS_O_WORKDIR variable gets 
> corrupted if its basename is identical to the name of the node.

I repeated your instructions and couldn't reproduce the problem.

Can you do 'qstat -f $jobid' to see if the server has the corrupted env vars?

Which scheduler are you using?  If using moab, kill it, submit the "bad" job,
look at 'qstat -f $jobid', start the scheduler, and look at 'qstat -f $jobid'
again and compare the difference.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://www.supercluster.org/pipermail/torqueusers/attachments/20070904/12d387f7/attachment.bin


More information about the torqueusers mailing list